J EAN -J ACQUES L ECERCLE Leaving the House of the Text: Deconstructing the Metaphor of Construction I Introduction The call for papers for this conference suggests that ‘construction’ and ‘constructivism’ are ‘basic analogies for epistemological and logical paradigms.’ I propose to extend the metaphor of construction to language (the sentence), to the text, and to examine its importance in the fields of linguistics and the philosophy of language, and to suggest ways of going beyond the metaphor, of leaving the house of the text, which might well be what Jameson, in another context, called “the prison-house of language.” 1 My point of departure is the adoption by the French language of the concept déconstruction, proposed by Derrida as what was at first a translation of the Heidegerrian term Destruktion. What puzzles me is why the translation he suggested was not the French word, destruction, why he had recourse to a word whose negative prefix presupposes the construction it seeks to negate in what sounds more like denial than negation. You will answer, of course, that this is the very movement of deconstruction: what it cancels, it retains under erasure. But perhaps there is more to Derrida’s choice than this. There is something radical in destruction: after the bomb, only rubble remains. Whereas deconstruction is slow and methodical, as in Hobbes’ famous paradox of the ship of Theseus: the ship rots bit by bit, and each plank is removed and replaced as it becomes rotten; but someone gets hold of each discarded plank in order to build himself a ship, or rather to rebuild the ship; at the end of the process there are two identical ships of Theseus, one brand-new, one rotten, which creates problems for the ascription of identity and for its stability over time. The ship has been at the same time deconstructed and reconstructed. But the real reason for the adoption of the term may well be that the metaphor of construction, as applied to either sentence or text, is inescapable, that it both frames and imprisons our thought about language and the text. The metaphor, of course, has a long history. It is not my intention here to chart its emergence in grammatical thought (where, etymologically, syntax has always been treated as a form of construction) and its contamination of 1 Fredric Jameson, The Prison House of Language (Princeton: Princeton UP, 1972). J EAN -J ACQUES L ECERCLE 198 the thought of the text, where it rivals the older metaphor of the text as cloth, derived from the etymological representation of the text as the result of a process of weaving. But what we call the moment of structuralism is certainly one of the moments when such contamination was most obvious. Then it was that the text became a prolonged sentence, supported by a structure, analogous to the syntactic structure of the sentence (and we remember that it is with Chomsky that syntax, formerly excluded from Saussurian langue and restricted to parole, became the core of the system of language). We find a literary illustration of this shift in a story by Donald Barthelme entitled “Sentence,” where the reader comes to realize that the apparent narrative is of no importance whatsoever, as the real protagonist is the sentence itself (and the story, which runs to several pages, is composed of one single sentence), a sentence that uses all the devices of embedding, parenthesis, and digression not to reach the final full stop that finally kills it. 2 If we wish to reconstruct the metaphor in order to deconstruct it and leave the house of the sentence and of the text, we must proceed in order, beginning with the sentence as construction. II The Sentence as Construction A dictionary of linguistic terms gives us two meanings of ‘construction’ in grammar. The first, called ‘traditional,’ refers to the way words are combined into sentences, according to meaning and syntactic function, in accordance with the grammatical rules of the language. The second, called ‘structural,’ states that a construction is any relevant combination of words or morphemes that enter into a larger construction. 3 Apart from the obvious circularity of such definition it is not in any way different from the definition of a syntagma, as opposed to paradigm, except that a syntagma fulfils a definite syntactic function, whereas the meaning of ‘construction’ is more vague, as the term can apply to almost any group of words, whether they constitute a proper syntagma or only a subpart of it. The most interesting aspect of the definition, however, concerns the shift from the traditional to the modern structural meaning of the term, as it illustrates the Chomskyan shift towards syntax as the core of language, with the use of an array of construction metaphors in his theoretical metalanguage: the syntactic trees that structure grammatical units are but visual representations of the step by stepconstruction of the sentence through the development of an algorithm of rules; the opposition between deep and surface structure gives the sentence 2 Donald Barthelme, “Sentence,“ Forty Stories, ed. Donald Barthelme (London: Secker & Warburg, 1988) 157-63. 3 Jean Dubois et al., Dictionnaire de linguistique (Paris: Larousse, 1973) 119. Leaving the House of the Text 199 the form of a construction, each syntactic transformation contributing to the final erection of the complete grammatical sentence; the multiplication of the modules that produce the building-blocks for the construction being an extension of the older dichotomy of the double articulation of linguistic signs, on which structural linguistics is founded. In the sentence, the scope of the construction metaphor is not limited to morphology and syntax. As the Italian word costrutto, meaning both grammatical construction and meaning as the result of a semantic construction indicates, the semantics of a language is also described through the construction metaphor. Here, a simple model of semantic construction will be represented by the treatment of the sentence as a Markov chain, that is as a series of finite states, represented by words, such that the choice of the first word of the sentence is entirely free, the choice of the second syntactically and semantically constrained by the first, and so on and so forth, till we reach the last word, the choice of which is totally constrained by the preceding choices. Thus, I can choose, in total freedom, to begin my sentence with the word ‘pride.’ A syntactic constraint (with limited choice) will induce me to choose a verb as a second word (the choice of the verb, however, is only limited by semantic constraints of compatibility with ‘pride’): The sentence is now ‘Pride comes… .’ ‘Come’ being an intransitive verb, I must choose for my third element an adverbial, not a noun phrase: so my sentence is ‘Pride comes before a… .’ You will not be surprised if I tell you that the last word is totally constrained, and the sentence must end with the noun ‘fall.’ The most interesting aspect of this theory of the syntactic cum semantic construction of the sentence is that, if it is not downright false, as Chomsky suggested in a celebrated essay, by pointing out the existence of syntactic operations of embedding which cannot be accounted for by recourse to a Markov chain, it is at least defeasible, the stylistic freedom of the speaker always allowing her to defeat the constraints, as when I choose not to end my sentence with the expected noun, ‘fall,’ but with another form of words, and my sentence now reads: ‘Pride come before a vote of impeachment.’ So I meant to tell you about Richard Nixon all the time, not merely to quote the most trite of proverbs. Here you have the essence of the Lacanian theory of the stitching point or upholstery button (point de capiton). The metaphor of the sentence as construction can be extended to the whole of language. Grammar is nothing but the construction of meaning through syntax. In the terms of Aristotle, it is a systematic passage, a path that is not yet a construction, from phone to logos. In the more modern terms of Jakobson, language works, in other words literally ‘makes’ sense (not only a praxis, but also a poiesis) along two axes: the paradigmatic axis, or axis of equivalence and selection, and the syntagmatic axis, the axis of combination; a vertical axis where metaphor thrives, versus a horizontal axis where me- J EAN -J ACQUES L ECERCLE 200 tonymies deploy. When the metaphor of construction comes to dominate, language is understood as a pyramid of layers which are themselves made of the building-blocks of language, phonemes, morphemes, words, and syntagmata. The construction of language takes the form of the piling up of phonetic, morphological, syntactic, and semantic levels. Aristotle’s horizontal path has become a vertical construction. III The Text as Construction, or the House of the Text The structural metaphor that allows us to move from the sentence to the text takes the form of a syllogism: a SENTENCE is a BUILDING, a TEXT is a PROLONGED SENTENCE, therefore a TEXT is a BUILDING. As is the case with the sentence, this metaphor is developed on several levels. The first level is grammatical or syntactic. Here the text is endowed with a structure in the architectural sense. Even as the glass skyscraper is held together by a metal structure, the text is kept together by the structuring of its narrative building-blocks. They can be fragments of text (clichés, scripts, quotations), but they can also be structural in a deeper sense, as in Greimas’ actantial model according to which a story is always the story of a gift and of a quest, as the sender gives the object to the receiver and the subject goes in search of the object: 4 Sender Object Receiver Helper Subject Opponent The six actants make up the abstract structure of the text: Their roles are filled by various actors (there may be more than one actor for a single actant, and a single actor may play the part of more than one actant). The structure of the six actants determines the combination and intersection of the narrative threads or sentences — indeed, in this model, the organic metaphor of weaving which still informs, for instance, Barthes’ reading model in S/ Z, 5 no longer holds, as the relationships between the various actants provide the sequences of narrative that are used as building-blocks. Thus, the first half of Dracula, in which the vampire is on the attack and travels to England where he vampirizes the heroines, Lucy and Mina, may be described as having the 4 Greimas’ semantic theories may be found in Algirdas J. Greimas, Sémantique structurale (Paris: Larousse, 1973). See also Ronald Schleifer, A.J. Greimas and the Nature of Meaning (London: Croom Helm, 1987). 5 Roland Barthes, S/ Z (Paris: Seuil, 1970). Leaving the House of the Text 201 structure of a quest (Dracula is the subject, the women are the objects) combined with a failed gift (nobody gives Dracula the right to take possession of the women), while the second half, in which the vampire is on the defensive and the leagued goodies travel to Transylvania in order to destroy him, may be described as a quest (where the goodies are the collective subject and the vampire the object), combined with a rightful gift (God, or the right that is on their side, authorise the goodies to hunt the vampire). But there is a second level of text structure: not the narrative/ syntactic level, but the deeper semantic level. Greimas’s actantial model is described by him as a surface structure, the deep structure being provided by basic logical-semantic oppositions between contraries, contradictories and subcontraries, as appears in the following diagram: 2. The Haunted A. Alive B. Dead 1. The Living 3. The Dead C. Not Dead D. Not Alive 4. The Undead The A,B,C,D squares represent the semantic oppositions between the two basic terms, ‘alive’ and ‘dead,’ that provide Dracula with its deep structure: A and B are contraries (there is the possibility of a third term, ‘between life and death,’ or ‘in the throes of death’); A and D or B and C are contradictories (there is no third term); A and C, or B and D are sub-contraries. Positions 1 to 4 indicate the realizations of the oppositions, taken two by two, in the novel: they distribute the characters into four categories (category 4 is occupied by the vampires — you have recognised their name in the novel: ‘the undead; ’ category 2 by those who have been bitten but are not quite dead yet). They also determine the narrative paths the characters take by opposing the normal horizontal path from life to death (1 to 3) to the unnatural vertical path taken by the vampirized, Lucy for instance (2 to 4). Mina Harker, the female heroine, who is vampirized by Dracula, escapes a fate worse than death, namely undeath, because of her unassailable virtue and married state. Therefore she goes from 1 to 2, from 2 to 4, and, fortunately, back to 1, which will enable her to proceed to 3, when her time comes, outside the compass of the novel, which ends on the announcement, or annunciation, of the birth of her first child. There is a third level of text structure, or text as construction: the argumentative level. It is at this level that the metaphor of construction comes into its own, as it provides not merely a figurative account of the structure of the J EAN -J ACQUES L ECERCLE 202 text, but constitutes that structure itself. I have already alluded to Lakoff and Johnson’s theory of structural metaphors in stating that the TEXT is a BUILD- ING. The metaphor of the building duly figures in Metaphors We Live By in the following form: THEORIES (and ARGUMENTS) ARE BUILDINGS Is that the foundation for your theory? The theory needs more support. The argument is shaky. We need some facts or the argument will fall apart. We need to construct a strong argument for that. I haven’t figured out what the form of the argument will be. Here are some more facts to shore up the theory. We need to buttress the theory with solid arguments. 6 We recognize the usual characteristics of structural metaphors according to that theory: They are not dependent on the materiality of language, but they form a system of set phrases; they intersect and combine with other structural metaphors. All we need to do to introduce the text is to combine the structural metaphor that is developed here with another one: the TEXT is an ARGUMENT; in other words, the text has an argumentative structure. That metaphor may not concern all texts, but it has sufficient generality to be of use to us (a narrative has the argumentative structure of a theory: It is regular, coherent, and methodical — playing with such structure, as in modernist or postmodernist texts does not cancel what it denies). What I am suggesting is merely the existence in many, perhaps in most texts, of a narrative or discursive structure which provides a frame for the construction of the text. It is time to sum up the characteristics of the construction metaphor in its two domains. We can do this through the following propositions: Proposition one: The text is a prolonged sentence. This in itself is yet another structural metaphor. It defines the moment of structuralism. Proposition two: Like the sentence, the text is composed of building blocks, whose combination, at their various levels, makes up its structure (the units may be narrative, discursive, actantial, semiotic, argumentative, or even philosophemes, if the text adopts the structure of a theory). A text is a kind of game of Lego. Proposition three: Language itself is a construction, made up of layers of structure, the building-blocks in each layer being ideally made up of the combination of building-blocks of the preceding layer (this is the principle of the double articulation of language). This proposition generalises, through analogy, the first two propositions: The metaphor allows us a va et vient, a going to and fro, between the local level of the text and the general level of language. 6 George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: U of Chicago P, 1980) 46. Leaving the House of the Text 203 Proposition four: Like language, and like the sentence, its actualisation, the text, is a layered structure (at the syntactic, semantic and narratological levels). This proposition combines the preceding two. It suggests that the text is a construction at two levels: the level of the elementary units, the buildingblocks, and the level (or rather levels) of the layers of building-blocks. Proposition five: The construction that makes up the text is its structure. Here the metaphor takes its truly architectural overtones; the text is not a mere piling up of layers of building blocks, it is a complex, because architectured, artefact. Proposition six: The consequence of the first five propositions is that the text is conceived of as an object, namely a building. This gives it a spatial existence (the text has clear spatial limits between the two covers of the book; metaphorically it has also narrative limits, namely a beginning, a middle, and an end, as in Aristotle’s famous definition of the complete story). And this gives it a temporal existence, which means that it is mortal but relatively stable (the degree of that stability in time, which can last for centuries, distinguishes the literary or philosophical text from others). The structure does not only provide internal coherence, but it also provides material coherence in space and time. Proposition seven: The spatial coherence of the text qua building makes it a unitary entity with a relevant solidarity of all its parts. The text is a totality and must be analysed as such. An explication de texte must account for the whole of the text, down to the semi-colons. And Lucien Goldman used to say that an interpretation of the text that leaves a remainder of unexplained aspects is a failure. Proposition eight: The temporal stability of the text explains not only the text’s survival in various historical conjunctures, and its capacity for recontextualisation (which is another specificity of the literary text), but can be extended to the internal coherence of the text in the form of its teleology: The text not only persists or endures in time, but it develops in time (the internal time of the text) as the actualization of a semantic or narrative programme, which means that the whole of the text is already contained in its very beginning (narratologists describe a narrative trope named the boucle of the text). Proposition nine: So far, we have neglected what is probably the philosophically most important characteristic of the metaphor: A building presupposes a builder, a structure in the architectural sense an architect. The metaphor, therefore, presupposes, in the matter of the sentence and of the text, a subject facing an object, or rather a subject creating or producing an object. And here the metaphor is not only structural but metaphysical: if the text qua construction is an object, a product, it can be isolated from the rest of phenomena, it can be handled and tampered with. It is the result of an activity of etymological poiesis, rather than mere praxis, in which language is the instru- J EAN -J ACQUES L ECERCLE 204 ment of the production process, and the text its result. Once we have reached these metaphysical heights, at the very top of the house, perhaps the time has come to take a jump and leave the house of the text. IV Leaving the House of the Text The last proposition has strongly suggested that the construction metaphor, in spite of its many advantages and obvious success, is linked to the metaphysics of subjectivity. This is a justification for the adoption of the term ‘deconstruction,’ in a philosophical tradition whose object is a critique of such metaphysics. We need only allude to Heidegger’s 1927 lectures on the fundamental problems of phenomenology, where he launches a critique of what he calls the metaphysics of production. This tradition would also concern interpretations of Marx such as the reading recently proposed by the French Marxist Franck Fischbach, where Spinoza, not Hegel, is seen as Marx’ philosophical predecessor. 7 In this Spinozan reading of Marx, man is an objective being, a part of nature, and emancipation does not take the form of an increased consciousness, but of man’s objectification. Conversely, consciousness and self consciousness are secondary effects of man’s objectification, in other words of the relation with the rest of nature that constitutes his essence. Alienation, in this view, is not the oppression of the conscious subject by a commodified world, but it is the focus on the individual conscious subject at the expense of man’s natural essence, an essence which is ‘generic,’ that is immediately social and collective rather than individual. According to Fischbach, Marx and Spinoza have three theses in common: 1) the secondary character of self consciousness; 2) the identity of nature and history; 3) an ontology of productive action. 8 The last thesis is of interest to us, not least because its formulation deliberately pastiches and inverts Habermas’ ontology of communicative action. This philosophical detour makes us aware of a contradiction within the construction metaphor. On the one hand, the text, being the result of productive action, acquires objectivity, and confers it to its producer (we are reminded here of the title of Macherey’s seminal book on the theory of literature, within another Spinozan tradition of reading of Marx, initiated by Althusser). 9 On the other hand, the metaphor insists on the instrumental quality of language, on authorial responsibility for the text, possibly through an intention of meaning, as an architectural structure is the consequence of a design. The ambiguity of the metaphor is embodied in the very term ‘struc- 7 Frank Fischbach, La production des hommes: Marx avec Spinoza (Paris: PUF, 2005). 8 Fischbach 29. 9 Pierre Macherey, Pour une théorie de la production littéraire (Paris: Maspero, 1966). Leaving the House of the Text 205 ture’: A structure can be natural, what scientific enquiry discovers in nature, independently of human action, as in the structure of a molecule; and it can be the result of human action and human planning, as in an architectural structure. This ambiguity in itself provides sufficient reason for abandoning the metaphor, leaving the prison house of language and the house of the text. In a sense, the science of language, as founded by Saussure, has always already left the house. The metaphors Saussure uses to help us think language through the new concept of langue are not metaphors of construction: a game of chess; the two sides of a single sheet of paper; language as a field of value, that is as a field of differences without fixed or substantial elements. The last metaphor is incompatible with the construction metaphor to the point of contradicting it: If there are only values, that is relations between elements that have no other existence except those relations (even as, in the Marxist doctrine, classes do not exist as independent entities, but only as the result of the class struggle), then there cannot be building blocks for the construction of either sentence or text, or rather such building-blocks (Saussure does not deny the existence of words) are only effects or actualizations of the field of value relations. We understand why Chomsky’s research programme, which presents itself as a continuation of Saussure, is a regression rather than a development because it relies heavily and almost explicitly on the construction metaphor. That Chomsky’s programme is a philosophical regression is well-known: It goes back to antiquated notions of innate ideas. And Chomsky translates the Saussurian dichotomy of langue and parole into his own dichotomy of competence and performance, which is supposed to play the same part in the system. But it does not: Whereas langue is collective and independent of the individual speaker, competence is situated in the individual speaker, in her genes or neurone circuits, so that speech is both maximally constrained (all creativity is rule governed, constrained by the innate apparatus of language, down to the smallest details, like the choice of reciprocal or reflexive pronouns) and located within an individual subject. Whereby Chomsky combines the gross materialism of biological determinism and the arrant idealism of methodological individualism. And his research programme also constitutes a grammatical regression in that, having abandoned langue as a collective entity separated from individual speakers (it is anterior and exterior to the speaker who has to assimilate it) and subject to historical change (whereas competence is situated in the arrested time of evolution), he resorts to the ancient metaphor of construction by treating the sentence as a construction made up of fixed building-blocks (the morphemes or words that combine into syntagmata), rather than as a meaningful path through a field of differential values. Chomsky’s syntagmatic markers, or grammatical trees, are not so much the actualisation of an algorithm of rules as blueprints for J EAN -J ACQUES L ECERCLE 206 the erection of a grammatical structure, the end product of which is the surface grammatical sentence. It is clear that in order to avoid the unwelcome philosophical and grammatical aspects of the Chomskyan research programme we must leave the house of language and the metaphor of the text or sentence as constructions. I shall briefly envisage one such escape with the help of the philosophy of Gilles Deleuze, in which we find series of concepts that allow us to think language outside the metaphor, even, Deleuze would argue, outside all metaphors. The reason why Deleuze cannot think language as a construction is that his ontology is an ontology of process and relations, of becomings, the actualization of the virtual, of transductivity, a concept he borrowed from Gilbert Simondon (a transductive relation is a relation which constitutes its terms instead of being posterior to them, and in an inductive or a deductive relation; Deleuze’s maxim is to always begin in the middle). 10 If you begin in the middle, in language and in the text, you will conceive it not as a construction, not as a tree but as a rhizome, a network rather than a structure, which can be entered at any point, in other words always in the middle. Rather than fixed units that can be used as building-blocks, you will have unstable elements in a state of continuous variation. Rather than a fixed structure that holds the text together, you will have a deployment on a plane of immanence. Rather than fixed rules that structure the structure, you will have regularities, agrammaticalities that make up a style, and lines of flight and movements of deterritorialization (in the field of language, such movements constitute a process of minorisation of the standard dialect). Rather than an individual speaker constructing her utterance or text, you will have assemblages: machinic assemblages of desire and collective assemblages of enunciation. In all this, Deleuze, who is explicitly hostile to Chomsky (to whom he tangentially assimilates linguistics as a whole, for instance in the fourth plateau of Mille plateaux), 11 is closer to Saussure than he thinks. So we have a series of concepts, plane of immanence, lines of flight, deterritorialisation, rhizome, variation, assemblages, agrammaticality (the rolling and pitching of language), and style that enable us to think language and the text without having to recourse to the metaphor of construction and its unwelcome philosophical grounding. I have tried elsewhere to detail the consequences that this philosophical language may have for an analysis of texts. 12 There are, however, two qualifications to be made here. First, the machinic in Deleuze is not the mechanical, so that a Deleuzian machine must not be treated as a construction: A machine is a constituent part of an assemblage, a 10 Gilbert Simondon, L’individu et sa genèse physico-biologique (Grenoble: Millon, 1995). 11 Gilles Deleuze and Félix Guattari, Mille plateaux: Capitalisme et Schizophénie (Paris: Minuit, 1980) 95-139. 12 Jean-Jacques Lecercle, Deleuze and Language (Basingstoke: Palgrave, 2002). Leaving the House of the Text 207 linking, which is not an articulation, of ontologically diverse elements. There is no builder of a machine, rather the human subject is an end product of the machinic assemblage, of which he is an integral part — no poiesis is involved. Second, we have left for good the realm of metaphor with the celebrated slogan: not metaphor but metamorphosis. The machine with its elements must be taken literally and in a constant state of flux, without stability or fixity. This is the view of language and of the text that can be derived from the common work of Deleuze and Guattari. The use of this series of concepts with regards to the text will be found in their Kafka and in Deleuze’s last collection of essays. 13 But there is an earlier version of Deleuze’s philosophy of language to be found in Logique du sens (and it is significant that in that book he seeks to think together language in general and the literary texts of Lewis Carroll). 14 Here, since the book is described in the preface as “an attempt at a logical and psychoanalytic novel,” 15 language is conceived as a series of Freudian modules, but not as layers or structure. The genesis of language in Logique du sens can be represented in the following table: Primary order Secondary organisation Tertiary arrangement Passion and Action of bodies Circulation of event Self, World and God: Individual as Person Scream Sense Doxa (good sense and common sense) Depth Surface Height Freud Lewis Carroll Plato Satire Humour Irony In the primary order, we find the violence of bodily affects, whose linguistic expression is the scream; in the tertiary arrangement, we have language as communication system, according to doxa, in other words good sense (the right direction of meaning) and common sense (a meaning shared with the community). In the middle we find the secondary organisation where Deleuzian events thrive: sense in this technical acceptation is not meaning, it logically and chronologically precedes it; it hovers over it as the 13 Gilles Deleuze and Félix Guattari, Kafka: pour une littérature mineure (Paris: Minuit, 1975); Gilles Deleuze, Critique et clinique (Paris: Minuit, 1993). 14 Gilles Deleuze, Logique du sens (Paris: Minuit, 1969). 15 Deleuze, Logique 7. J EAN -J ACQUES L ECERCLE 208 event of the battle hovers over the battle like a mist. Sense keeps on the surface of language, in the paradoxes and portmanteau words of which Lewis Carroll was so fond of, it avoids the depths of the Freudian primary order, and rejects the heights of Socratic irony: Its domain is the domain of humour, as opposed to violent satire and patronizing irony. This system of metaphors, and the concept, ‘sense’ that emerges out of it, is what enables us to leave the house of the language, to think about language outside the construction metaphor. V Conclusion The path I have followed (I dare not say the argument I have constructed) can be summarised by the following propositions. Proposition one: The construction metaphor, as applied to sentence and text, is part of our common sense (hence its proximity to the theory of metaphor advocated by Lakoff and Johnson). In the thinking of language, it is of considerable antiquity. Not quite so with regards to the text (here, the moment of structuralism is of particular importance). Its explanatory advantages are not in doubt. Proposition two: Its philosophical background is characterised by the tension between an ontology of production and a philosophy of the selfconscious subject, a philosophy which adopts as its first concepts a subject and an object facing each other, with precedence being given to the subject. Proposition three: In a sense, linguistics as the science of language has always already avoided the metaphor, which plays no part in Saussure’s conceptual system. Proposition four: But the metaphor has invaded linguistics with the Chomskyan research programme, which is heavily dependent on ancient grammar (in the shape of the parts of speech that provide the building-blocks for the construction of the sentence). The Chomskyan programme is therefore a regression, both in its philosophy of language and in its linguistic theory. Proposition five: Contemporary philosophy (and philosophy of language) provides various routes of escape from the metaphor. I have evoked three, deconstruction, a renewed Spinozan interpretation of Marx, and Deleuze’s conceptions of language. At the end of this argumentative path, I hope that the philosophical and linguistic advantages for leaving the house of the text are equally not in doubt. And I hope to have answered two questions. One, this was my starting point, why did Derrida choose ‘deconstruction’ as the name of his critical practice? The answer is: Because the term indicates what it does: in order to think language, and the text, we must go through the metaphor of construc- Leaving the House of the Text 209 tion, and we must leave it (hence the negative prefix). And since the practice may be generalised, from thinking language to thinking tout court, I think I have answered a second question, which I never asked: Why are poststructuralist philosophers like Derrida and Deleuze so fascinated with, and hostile to, metaphor (in the case of Deleuze, the hostility is explicit)? Because, in order to think, we must go through the common sense that systems of metaphors, in the sense of Lakoff and Johnson, embody, and we must leave it in order to reach the realm of the philosophical concept, which, going against the grain of doxa, of common sense, has no truck with metaphor and involves a different type of systematicity. I shall leave the last word to Gilles Deleuze, by repeating his slogan: ‘Not metaphor, but metamorphosis! ’ J EAN -J ACQUES L ECERCLE 210 Works Cited Barthelme, Donald. “Sentence.” Forty Stories. Ed. Donald Barthelme. London: Secker & Warburg, 1988. Barthes, Roland. S/ Z. Paris: Seuil, 1970. Deleuze, Gilles. Logique du sens. Paris: Minuit, 1969. ---. Critique et clinique. Paris: Minuit, 1993. Deleuze, Gilles, and Félix Guattari. Mille plateaux: Capitalisme et Schizophrénie. Paris: Minuit, 1980. ---. Kafka: Pour une littérature mineure. Paris: Minuit, 1993. Dubois, Jean, et al. Dictionnaire de linguistique. Paris: Larousse, 1973. Fischbach, Frank. La production des hommes: Marx avec Spinoza. Paris: PUF, 2005. Greimas, Algirdas Julien. Sémantique structurale. Paris: Larousse, 1973. Jameson, Fredric. The Prison House of Language. Princeton: Princeton UP, 1972. Lakoff, George, and Mark Johnson. Metaphors We Live By. Chicago: U of Chicago P, 1980. Lecercle, Jean-Jacques. Deleuze and Language. Basingstoke: Palgrave, 2002. Macherey, Pierre. Pour une théorie de la production littéraire. Paris: Maspero, 1966. Schleifer, Ronald. A.J. Greimas and the Nature of Meaning. London: Croom Helm, 1987. Simondon, Gilbert. L’individu et sa genèse physico-biologique. Grenoble: Millon, 1995.